The present invention relates to the processing of image data, wherein said data takes the form of a plurality of image pixels.
Many procedures for the modification of image data after the data has been recorded are known and are collectively referred to as post production. It is known to use analogue circuitry for modifying television and video signals, in which said signals may be represented as red, green and blue components or as luminance plus chrominance components. When working with video signals, part of the information may be removed or keyed out at particular times or positions within the image, defined by a synchronised key, matte or alpha signal. Alternatively, parts of the video signal may be suppressed to black in response to a suppression signal. These keying or suppression signals may be derived from part of the video image itself, often with reference to particular colours within the image in accordance with a process generally referred to as chroma-keying.
It is also known to modify digital representations of video images derived by sampling an image to produce an array of picture elements (pixels) in which each of said pixels may represent a colour defined by three colour components stored as three numerical values. Thus, it is known to represent each of red, green and blue colour components by respective eight-bit words representing the colour of the pixel at particular positions within the image array.
When chroma-keying techniques are to be employed, it is preferable for the production process to record the information in such a way as to facilitate the chroma-keying process. Thus, it is preferable for foreground talent to be recorded against a saturated blue or green background for example, thereby allowing any portions of the image which are perceived as being saturated blue or green as belonging to part of the keying signal. However, in many situations where it would be desirable to employ post-production techniques, the original source material will have been recorded under less than favourable conditions. Under these circumstances, it can become difficult to accurately define those colours which are to constitute part of the full key, those colours which are to constitute part of the full key-off and those colours which are used to generate a blending between the two extremes, in an area which may be considered as the key softness.
In many situations it may be possible to define a first set of colours for full key-on in combination with a second set of colours for full key-off. However, an appropriate interface has to be provided to a user to enable a selection to be made and in many known environments, although distinctive sets of colours do exist, it is possible that the controls available to an operator would not be sufficiently sophisticated in order for these regions to be clearly defined. In these circumstances, it is known for the key to be generated in several passes or for part of the image to be dealt with in a different way, possibly by the generation of an appropriate garbage mask. All of these known approaches tend to increase the expertise and time required in order to generate an appropriate chroma-key, thereby significantly limiting the availability of the technique.
A system in which each pixel is represented by three colour components defining a position within colour-space is disclosed in International patent publication WO 98/11510. A base colour is identified and a distance in colour-space between an input colour and said base colour is calculated. The control value, which may be used as a basis for a chroma suppress or a chroma-key etc, is produced in response to the calculated distance.
The system disclosed in the aforesaid patent publication provides improved procedures for producing a chroma-key thereby achieving improved chroma-keying results. However, the problem with the system disclosed in this publication is that it is difficult for users to control operational parameters so as to make optimum use of the processing facilities available.
According to a first aspect of the present invention there is provided a method of processing image data wherein said data takes the form of a plurality of image pixels, comprising steps of displaying a three dimensional representation of colour-space; analysing the colour of pixels within said image; and displaying said analysed pixels at colour related positions within said displayed colour-space.
In a preferred embodiment, the pixels have a unifying property and this property may be that the pixels belong to a background area which is to be replaced by a new background area. The background area may be selected by viewing an image, manually adjusting a cursor in response to activation of an interface device and selecting example pixels of the background image.
In a preferred embodiment, the method further comprises the steps of determining the position of a volume in colour-space that contains the displayed pixels and displaying said bounding volume. The bounding volume may be a cuboid, a minimised convex hull or an ellipsoid. The geometry of the displayed bounding volume may be adjusted by manual operation of an interface device.
In a preferred embodiment, pixels displayed in colour-space are displayed substantially in their original colour.
According to a second aspect of the present invention, there is provided image processing apparatus configured to process data taking the form of a plurality of image pixels, comprising display means configured to display a three-dimensional representation of colour-space; analysing means configured to analyse the colour of pixels within said image; and processing means configured to supply analysed pixels to said display means at colour related positions within said displayed colour-space.